Frontier AI is getting more capable, and much harder to govern.
OpenAI says Astra can independently find and exploit zero-day vulnerabilities. Anthropic is loosening restrictions on its newest models while tightening enterprise safeguards. Apple is escalating its legal fight with OpenAI. And governments are now debating how much regulation is too much.
The 5 things worth knowing today:
OpenAI says Astra can independently find and exploit zero-day vulnerabilities
Anthropic launched cheaper and less restrictive Fable 5.1
Apple is escalating its trade secrets fight with OpenAI
The U.S. is pushing G20 countries toward lighter AI regulation
Thrive Capital’s proposed $4.2B FIFA deal collapsed amid legal scrutiny
FROM OUR PARTNER - KAPITALWISE
💹 Your financial goals are personal. Your advisor should be too.
Not every financial advisor is right for every investor. The right fit depends on what you’re building toward, the kind of guidance you need, and where you are financially.
Kapitalwise AdvisorConnect uses AI-powered matching to connect you with vetted financial advisors and planners based on your needs - making it easier to find relevant expertise without spending hours searching and comparing.
Think of it as a smarter starting point for an important financial decision: fewer random options, more relevant matches.
STARTUPS RAISING MILLIONS
💰 Startup Funding Updates
At least $937.45M + €27M + CAD 2.65M disclosed across today’s startup rounds
The three biggest rounds:
Tripo AI: $445M Series B
Scan : $220M
Odyssey Energy Solutions: $74M
Together, those three companies accounted for nearly 79% of today’s disclosed dollar funding.
PRE-SEED
Visko: Sunnyvale-based AI research company developing real-time interactive world models raised $10M led by Llama Ventures.
DataAgent: Tel Aviv-based developer of a remediation-first Digital Immune System for modern applications raised $10M led by MizMaa Ventures and Alicorn Venture Partners.
Scopia Surgical: Montreal-based TechMed company developing real-time AI-powered surgical navigation for minimally invasive and robotic surgery raised CAD 2.65M co-led by Linearis Ventures and Anges Québec, with participation from Fonds Impulsion, HaloHealth, Richard Rubin, Suresh Madan, and others.
SEED
Aranya: San Francisco-based AI-native multicluster operating system for GPU infrastructure raised $11M across seed and pre-seed financing. The latest $9M seed round was led by First Round Capital, with participation from BoxGroup, Vermilion Cliffs, and Asylum Ventures.
Inleap Photonics: Hanover-based deeptech photonics startup raised approximately €20M in seed funding led by UVC Partners, with participation from HTGF, Balnord, Sentris, and Scale Capital.
Kepler Aerospace: Bengaluru-based space infrastructure and autonomous mission operations startup raised $8M led by Blue Ashva Capital, with participation from Finvolve and India Accelerator.
Senticell: Philadelphia-based biotechnology company developing a new class of liquid biopsy using red blood cell-bound nucleic acids raised $7M from The Venture City, Capital Factory, Scalpel Ventures, Evoce Capital, and others.
Guickly: San Jose-based AI measurement platform for enterprises raised $4.2M led by Engineering Capital, with participation from Converge VC, Neon Fund, and angel investors.
TrustedRouter: Miami-based AI infrastructure and API orchestration startup raised $1.25M from Slow Ventures, Bill Tai, Linda Avey, and George Xing.
GROWTH
Tripo AI: Beijing-based 3D-native foundation model platform for generative 3D raised approximately $445M Series B led by MPCi, with participation from Perfect World, BlueFocus, SPC, Yanqu Games, ThunderSoft, 37 Interactive Entertainment, CDH Venture and Growth Capital, CICC, CMC Capital Partners, Primavera Capital, INCE Capital, and others.
Scan.com: Atlanta-based medical imaging network platform raised $220M in combined equity and debt financing. The package included $90M in equity led by Noteus Partners, with participation from Aviva, Concord Health Partners, YZR Capital, Oxford Capital, and others, plus $130M in debt from VerisFi Capital and Atempo Growth.
Odyssey Energy Solutions: Boulder-based financing and procurement platform for distributed renewable energy projects raised $74M, consisting of $27M in equity and $47M in debt. The round was backed by Broadscale Group, FMO, Al Mada Ventures, Union Square Ventures, Equal Ventures, Abstract Ventures, FJ Labs, MCJ, and others.
Light: Houston-based retail energy software and grid orchestration platform raised $46M Series A led by Matrix, with participation from Activate Capital, Spark Capital, Mischief, Gigascale Capital, MCJ, and BoxGroup.
Reframe Systems: Boston-based physical AI company industrializing home construction raised $40M led by Energy Impact Partners, with participation from Counterpart Ventures, E12 Ventures, Global Brain, Thin Line Capital, Up Partners, LACI Impact Fund, Eclipse, VoLo Earth, Cubit Capital, RA Capital Management, MassMutual Catalyst, and Nor’easter.
Gridsight: Sydney-based AI-powered capacity management platform for utilities raised $26M Series B led by Insight Partners, with participation from Galvanize, Airtree, Energy Transition Ventures, and Aera VC.
Clipto: San Francisco-based AI video search and content discovery startup raised $15M at a $250M valuation from HSG, GL Ventures, EnvisionX Capital, Palm Drive Capital, Hans Tung, Lu Zhang, and 522 Ventures.
Norbert Health: New York-based physical AI control system for robotic care platforms raised $14M Series A backed by William A. Marino with Cardinal Group, with participation from Exor Seeds, CareIT, Alexis Le-Quoc, Olivier Pomel, Saeju Jeong, and Thomas Clozel.
Blue Voice: Boston-based public safety software and legal technology startup raised $6M jointly led by SignalFire and Las Olas Venture Capital.
Xorlab: Zurich-based cybersecurity and enterprise email defense startup raised €5M Series A+ led by Spicehaus Partners, with participation from Grapha Holding, EquityPitcher Ventures, and ZKB Start-up Finance.
Holifya: Milan-based digital clinic using AI and advanced diagnostics for obesity treatment raised €2M, consisting of €1.6M in equity plus €0.4M in debt capital and grants. The round was led by Next Age, with participation from Doorway, 20Fund, Zooga VC, and several angel investors.
📍 Funding Signal
AI is moving deeper into infrastructure and the physical world.
Today’s funding went far beyond software, with capital flowing into GPU infrastructure, generative 3D, robotic healthcare, energy grids, automated construction, quantum sensing, space infrastructure, and medical imaging.
Tripo AI alone raised $445M, while companies like Light, Reframe Systems, Gridsight, and Norbert Health show how AI is being embedded into real-world industries.
Another notable trend is blended financing. Scan.com, Odyssey Energy Solutions, and Holifya combined equity with debt or grants, showing how capital-intensive startups are using more flexible funding structures as they scale.
NEW VCs IN THE MARKET
🏦 Venture Capital Updates.
Thrive Capital’s $4.2B FIFA deal collapses amid legal scrutiny
Thrive Capital is facing scrutiny over a proposed $4.2B investment in FIFA Forward Enterprise, a planned vehicle that would have commercialized parts of FIFA’s media and ticketing revenues at a reported $20B valuation.
The deal was ultimately scrapped after opposition from European and North American soccer associations. Thrive has since hired attorney Alex Spiro to represent the firm as the dispute moves into a legal phase.
📍 Capital signal
Large venture and growth investors are increasingly looking beyond traditional technology companies for outsized opportunities, but complex governance, stakeholder alignment, and regulatory risk can become just as important as valuation and growth potential.
KEY STORIES IN TECH
📜 Latest In Tech
Anthropic launches cheaper, less restrictive Fable 5.1
Anthropic released Fable 5.1 and Mythos 5.1, with Fable bringing lower token costs, fewer false-positive safety restrictions, and stronger benchmark performance.
Fable 5.1 is available through Anthropic’s API and cloud platforms, while Mythos 5.1 remains limited to approved cybersecurity and life sciences partners.
Anthropic is also expanding zero data retention and enterprise safeguards, allowing customers to run models with tighter control over sensitive data and monitoring.
Why it matters: Anthropic is trying to make its most capable models easier and cheaper to deploy in enterprise environments without giving up safety controls, which could strengthen its position against OpenAI and Google in high-value business use cases.
OpenAI’s Astra can find and exploit zero-day vulnerabilities
OpenAI says its upcoming Astra model is the first of its models to cross the company’s “critical cybersecurity threshold,” meaning it can discover unknown software vulnerabilities and exploit them without human guidance.
Astra reportedly scored perfectly on ExploitBench and found two zero-day vulnerabilities in a modified internal test, prompting OpenAI to limit access to its most advanced cyber capabilities.
OpenAI says it is adding stricter abuse detection, higher-risk account restrictions, jailbreak protections, and chain-of-thought monitoring before the model is released.
Why it matters: Frontier AI models are moving from helping with cybersecurity analysis to independently finding and exploiting vulnerabilities. That raises the stakes around model access, safety controls, and how quickly defensive systems can keep up.
Apple accuses OpenAI of destroying evidence
Apple says a forensic review of former engineer Chang Liu’s laptop uncovered evidence that confidential circuit designs were used after he joined OpenAI.
The company also alleges Liu asked an OpenAI colleague to destroy evidence after learning about Apple’s internal investigation.
Apple is now seeking access to additional hardware, including a Mac mini that allegedly synced the confidential file through iCloud.
Why it matters: If Apple can prove its hardware trade secrets were used inside OpenAI, the case could create serious legal and operational risks for OpenAI’s expanding hardware ambitions.
U.S. pushes G20 toward lighter AI regulation
The U.S. is urging G20 countries to avoid creating broad new AI regulatory bodies and instead focus rules only on genuinely new risks or situations.
China signed on to the proposed “Carolina Principles,” while major tech leaders including Mark Zuckerberg, Elon Musk, Demis Hassabis, Sam Altman and Jensen Huang are participating in the broader discussions.
The push broadly aligns with the interests of leading U.S. AI companies, which argue that heavy regulation could slow model releases, raise compliance costs and weaken competitiveness against China.
Why it matters: AI regulation is becoming part of the geopolitical race for technological leadership. If more countries adopt a lighter-touch framework, U.S. AI companies could gain more room to scale globally while shaping the rules around frontier models.
AI labs are buying tens of thousands of Macs to train agents
AI companies including OpenAI are reportedly buying tens of thousands of Mac minis and Mac Studios to train and test computer-use agents through reinforcement learning.
Apple’s unified memory architecture makes Macs useful for repeatedly running agents inside operating-system environments, while Anthropic is also accessing Mac mini capacity through AWS.
Demand has become strong enough to stretch delivery times for high-memory Macs and push some buyers toward alternatives like Nvidia’s DGX Spark.
Why it matters: The AI infrastructure boom is expanding beyond traditional data-center GPUs. As computer-use agents grow, Apple could become an unexpected beneficiary by supplying the hardware needed to train and test them.
Anthropic resumes cyber testing after Claude escaped test environments
Anthropic restarted external cybersecurity evaluations after pausing them following three incidents where Claude models reached real-world systems instead of staying inside simulated targets.
In one case, Claude Opus 4.7 reportedly attacked a real company that shared a domain name with a fictional target, while another test leaked Python code onto the public internet.
Anthropic traced the failures to a misconfigured sandbox from evaluation partner Irregular that did not properly block internet access.
Why it matters: As AI models gain stronger offensive cyber capabilities, testing infrastructure itself is becoming a critical safety layer. These incidents show that evaluation failures can create real-world consequences before the models are even widely deployed.
LAST COFFEE SIP
🌎 Around The Ecosystem
Apple: John Ternus has officially taken over as CEO after 25 years at Apple, rising from the product design team to lead hardware engineering across the iPhone, Mac, AirPods, Apple Watch and Vision Pro before succeeding Tim Cook.
Amazon + FTC: The FTC and attorneys general from 21 states sued Amazon over alleged hidden increases in minimum ad prices, claiming the company illegally collected more than $20B from advertisers over seven years.
Dyson: Launched the $499 CameraJet toothbrush with a built-in 100,000-pixel camera that identifies gaps between teeth and automatically targets them with a mouthrinse jet, using a vision system trained on 470,000 dental images.
MUST READ
📚 Worth Your Time
The reference calls that decide your round: Most founders prepare for the pitch, but not the calls investors make after the meeting. This piece shows how VC reference checks really work, from former colleagues and ex-cofounders to passing investors, customers, and off-list backchannels.
The YC seed deck structure: Most founders build decks from random templates. This piece breaks down the repeatable structure used by YC startups that raised $450M+. It also covers the simple design rules that make a deck clear, readable, and easy for investors to understand quickly.
HIRING ALERT: STARTUPS & VC ROLES
💼 Today’s VC & Startup Job Opportunities
Associate, Healthcare | B Capital Group | USA | Apply Here
Portfolio Reporting Team Lead | Cerity Partners | USA | Apply Here
Associate | Mount Judi Ventures | USA | Apply Here
Technical Investment Operations Associate | Alpaca VC | USA | Apply Here
VC Research Lead | Murph Capital | Remote | Apply Here
Trade Operations Analyst | RA Capital | USA | Apply Here
Portfolio Management Associate | Cerity Partners | USA | Apply Here
Corporate Controller | a16z | USA | Apply Here
Investment Associate | Cerity Partners | USA | Apply Here
Trying to break into venture capital?
And if you’re serious about building a career in venture capital, subscribe to BREAK INTO VC newsletter. We regularly share VC jobs, career resources, investing frameworks, and practical insights to help you break into and grow within venture. It’s written by Ex-VC.
PARTNERSHIP WITH US
🤝 Get your brand in front of 125,000+ decision makers.
Our newsletter is read by 125,000+ of founders, CEOs, investors (VCs / Angel Investors) and senior operators. Get in touch today.
Do you have any news to share about your Startup or VC fund? Please email us.
We’ll be back in your inbox on Tomorrow.